Probabilistically Checkable Proofs. 1 Introduction to Probabilistically Checkable Proofs
|
|
- Judith Lang
- 5 years ago
- Views:
Transcription
1 Course Proofs and Computers, JASS 06 Probabilistically Checkable Proofs Lukas Bulwahn May 21, Introduction to Probabilistically Checkable Proofs 1.1 History of Inapproximability Results Before introducing probabilistically checkable proofs, I shortly give an overview of the historical development in the field of inapproximability results which are closely related to PCPs. A foundational paper from Johnson in 1974 states approximation algorithms and inapproximability results for Max SAT, Set Cover, Independent Set, and Coloring. While the decision problems for various problems, such as Max SAT, Set Cover, Independent Set, and Coloring, were shown to be NP-hard by Cook, Levin and Karp, it was difficult to show approximability and inapproximability results with the known reductions. To prove inapproximability results, there was a new model for NP necessary, this evolved from works on multi-provers interactive proofs from Ben-Or and Goldwasser. In 1991, Feige and Goldwasser created this new model and showed new inapproximability results with this model. This new model was later on called PCP. In 1992, Arora [?] could prove the PCP Theorem which made PCP easier and useful to apply for many inapproximability problems. Since then, many computer scientists could prove and improve many inapproximability results creating tight results for many NP-hard problems. In 2005, Dinur has published a new proof for the PCP Theorem. This will be introduced in the paper from Bernhard Vesenmeyer, the necessary tools for this proof will be introduced in the sections about constraint graphs and expander graphs. 1
2 1.2 A (r,q)-restricted verifier First, we have to define the verifier which will prove for an input x to be part of the language or not. Definition 1. A verifier V is a (r,q)-restricted verifier if for any input x, witness w, and random string τ of length O(r), the decision V w (x, τ) = yes is based on at most O(q) bits from the witness w. This verifier can also be seen as an interactive proof system, but in this interactive proof system the number of random bits of the verifier is restricted. In the definition of the (r,q)-restricted verifier, you see that we will consider the number of random bits we can use (randomness complexity) and the number of queries to the witness (query complexity). Specifying the random and query complexity results in different classes for languages as we will see in the examples later. A (r,q)-restricted verifier is called non-adaptive if the queries to the witness w only depend on the input x and the random string τ. If the next queries are also dependant from the previous queries from the witness w, the verifier is called adaptive. For the sake of simplicity we will consider a verifier non-adaptive from now on. 1.3 Probabilistically checkable proofs Definition 2. A language L is probabilistically checkable using an (r,q)-restricted verifier V iff Completeness: If x L then there exists a witness w such that P r τ [V w (x, τ) = yes ] = 1. Soundness: If x / L then for every witness w we have P r τ [V w (x, τ) = yes ] < 1/2. Comparing this with the definition of interactive proof systems, we also see the two-sided error is reduced to an one-sided error up to 1 2 for an input which 2
3 is assumed to be not accepted. It is clear that the boundary of 1 2 is mostly arbitary, as you can run many (up to any constant c) proofs which can reduce the probability of failure to any ɛ > 0. However, the number of queries is still interesting, since for inapproximability results this creates the specific gap of problem. Today, it is known that 11 bits are enough and there are still efforts to lower the number. Example 3. Some simple examples for PCP-Classes are: P = PCP(0, 0) : This verifier has no witness and no randomness. Therefore it can only work on the input x, just like a deterministic polynomialtime bound Turing machine. NP = PCP(0, poly) : This verifier has no randomness and a witness which can be read at polynomial many positions, this is identical to a NP witness. Without the randomness, acceptance is inprobabilistic. The verifier is the same as a non-deterministic polynomial-time bound Turing machine. NP PCP(log, poly) : From the example above, it is simple to see that by allowing logarithmical many random bits one will not reduce the computation of the verifier. co-rp = PCP(poly, 0) : The verifier has no witness and can use random bits in each step. The probability of false positive is less than 1/2. This equals the definition of co RP. These examples are simple because at least one of the complexities is set to zero. More interesting is how increasing one of the complexities and decreasing the other changes the power of the verifier. This will finally lead to the PCP Theorem. 2 PCP Theorem - Part 1 First notice that it can be easily shown that PCP(log,poly) NP. So it is clear that NP = PCP(log,poly). The amazing fact is now that it can also be shown that: 2.1 PCP Theorem Definition 4 (PCP Theorem). NP = P CP (log(n), 1) The PCP Theorem states that the verifier only has to look at a constant number of bits of the witness string in order to probabilistically say if the input is in the language or not. 3
4 First, we will prove the easier side of the PCP theorem: PCP(log(n), 1) NP. The main idea of the following proof is that a non-deterministic Turing machine V can create all random strings with logarithmic length of a verifier V and simulate the calculation with the input x and all created random strings in polynomial time. The first step is to explain how the (log(n), 1)-verifier V works with the input x. Proof. Let L PCP(log(n), 1) there is a (log(n), 1)-verifier V. For τ there are 2 O(log(n)) n c many random strings, namely τ 1,, τ nc. The verifier V will work as follows: 1. Reads a random string τ i, 1 i n c. 2. Uses x and τ i to calculate q positions i 1,, i q to read from the witness string. 3. Run a calculation with x and w i1,, w iq, and answer yes or no. Proof. Now, we will simulate the verifier V on a non-deterministic Turing machine V. The witness string for V is w which has polynomial length since V can only access polynomial positions. V now calculates step 2 and 3 from V for every possible τ i and answers yes if all simulated calculations of V answered yes. It is left to show that V behaves like V. Proof. x L and L P CP (log(n), 1) For a given w, V returns yes with probability 1. With this witness w V will also return yes. x / L There is no witness string w for V because at least on half of the calculations will not answer yes. As the verifier V can simulates all possible random runs of V, one would think at first glance that NP is more powerful than PCP(log(n), 1). Showing the inclusion of NP PCP(log(n), 1) is one of the most difficult proofs in theory of computation - an overview of the original proof can be found in [?] - and therefore we will first show how to apply the PCP Theorem brings great results in inapproximability. 3 Applying PCP Theorem 3.1 Selected parts of approximability Simply speaking, an approximation problem is an problem which one is not interested in finding the best possible solution, but in finding a solution which 4
5 is close to best solution and its computation is still efficient. The following definitions will formally describe the problem: Definition 5. An optimization problem O is defined by a cost function C : Σ Σ R + { }, that given an instance string x and a solution string s outputs C(x, s) which is either the cost of the solution or if the solution is illegal. Let OP T (x) denote the optimal value a solution can get, then: OP T (x) = max s:c(x,s) C(x, s). An optimization problem is to find a legal solution s that attains the optimal value of cost, C(x, s ) = OP T (x). Example 6. MAX-3SAT is the problem of finding an assignment A which maximizes the percent of satisfied clauses of a 3CNF formula ψ. Of course, if ψ is satisfiable, then the optimal value of MAX-3SAT is 1. Definition 7. A is an r-approximation algorithm for a maximation problem iff for any input x, A finds a solution s that C(x, s) rop T (x). In this paper, we are using the definition that 0 < r < 1 and the better an approximation is, the higher r is. In other literature, you also can find the definition for r = 1/r and r > 1, then the better an approximation is, the lower r is. It is clear that reducing a decision problem is easier than reducing an r-approximation problem, so we introduce the definition for gap problems: Definition 8. Let O be a maximization problem. Let x be an instance of the problem. A gap(α, β)-o is the problem of deciding between the following alternatives: Yes : OP T (x) β No : OP T (x) α If OP T [α, β) then both alternatives are acceptable. Intuitively, A gap problem is to split all possible inputs into the ones for which the optimal solution is less than α and the others for which the optimal solution is greater than β. If the gap problem is NP-hard, then it is the α β - approximation algorithm is also NP-hard. Now, reductions from gap problems can be used instead of working directly with r-approximation problems. 3.2 Equivalence of PCP Theorem and gap-max-3sat is NP-hard. In this section, we will reduce gap-max-3sat to 3SAT using the PCP Theorem. There can also be found many other reductions using the PCP Theorem, e.g. 5
6 for the clique number or the chromatic number of graphs in [?]. Lemma 9. The following statements are equivalent: 1. (PCP Theorem) NP = P CP (log(n), 1) 2. There exists α (0, 1), such that gap(α, 1)-MAX-3SAT is NP hard. Proof. (2 1) Let language L NP. Assumption: gap(α, 1)-MAX- 3SAT is NP hard. = there exists a 3CNF formula, ψ x,l = c 1... c m, such that 1. x L ψ x,l is satisfiable. 2. x / L for every assignment A, the number of clauses in ψ x,l that are satisfied is less than αm. The verifier V can use the following algorithm to check if a string x is in the language L: Algorithm 1. step: Construct the 3CNF formula ψ x,l. 2. step: Get an assignment A and create witness/proof w = ψ x,l A. 3. step: Choose k = O(1) clauses from the witness. 4. step: If all k clauses are satisfied, return yes. Proof. Completeness: If the assignment A satisfies the formula, V will answer yes no matter which k clauses were chosen. Soundness: If A does not satisfy ψ x,l, then it satisfies at most αm clauses = the probability to answer yes is at most α k. With k > log(1/2)/log(α) : P r τ [V w x, τ = yes ] α k = P r τ [V w (x, τ) = yes ] < 1/2. Proof. (1 2) Proof by reduction from gap-max-3sat to 3SAT. 3SAT NP = 3SAT P CP [log, 1] = there exists a verifier V such that a given 3CNF formula φ: φ is satisfiable w : P r τ [V w (φ, τ) = yes ] = 1. φ is not satisfiable w : P r τ [V w (φ, τ) = yes ] < 1/2. 6
7 Proof. The verifier only considers q bits of the witness w for its decision. acceptance is determined with local constraint ψ φ τ and variable assignment according to the positions in the witness w. It is still true that: φ is satisfiable all local constraints ψ φ τ are satisfied. φ is not satisfiable for any assignment A at most half of the local constraints are satisfied. Proof. Construct a new formula φ = ψ φ τ 1 ψ φ τ c n with τ 1, τ n c are all random string of the length O(log(n)). For φ, we have: φ is satisfiable φ is satisfied. φ is not satisfiable any assignment for φ satisfies at most half of the clauses of φ. Proof. Construct a 3CNF formula from each local constraint. φ = (ψ 1,1... ψ 1,k )... }{{} ψτ φ 1 (ψ nc,1... ψ nc,k) }{{} ψ φ τ n c φ is satisfiable φ is satisfied. φ is not satisfiable any assignment for φ satisfies at most 2k 1 2k clauses of φ. of the This concludes the reduction from gap(α,1)-max-3sat for α = 2k 1 2k assuming the PCP Theorem. to 3SAT 3.3 Outlook of MAX-3SAT After we have shown that there exists an α, it would be interesting for which α this was already shown. Theorem 10 (John Hastad, 1997). For any α ( 7 8, 1), the problem gap(α,1)- MAX-3SAT is NP-hard. The proof for this theorem can be found in [?]. Fact But notice this interesting fact: Howard Karloff and Uri Zwick have stated a 7 8-Approximation Algorithm for MAX-3-SAT and provided strong evidence that the algorithm performs equally well on arbitrary MAX-3-SAT instances. 7
8 So for this problem, there has been found a best possible performing algorithm. The approximation algorithm and the inapproximability result have created a clear boundary for MAX-3-SAT. For other problems, research are still going on. To become an overview of the results, one can find tables in [?], [?] and [?]. 4 PCP Theorem - Part 2 The following presentation [?] will present the proof for the PCP Theorem by Gap Amplification following the proof of Irit Dinur [?]. This proof uses constraint graphs and expander graphs. Therefore, we shortly will introduce these two topics. 4.1 Constraint Graphs Definition 11. G = (V, E), Σ, C is called a constraint graph, if 1. (V,E) is an undirected graph, called the underlying graph of G. 2. The set V is also viewed as a set of variables assuming values over alphabet Σ. 3. Each edge e E carries a constraint c e : Σ 2 {T, F } and C = {c e } e E. Definition 12. An assignment is a mapping σ : V Σ that gives each vertex in V a value from Σ. For any assignment σ, define SAT σ (G) = P r(c e (σ(u), σ(v)) = T ) and SAT (G) = max σ SAT σ (G). Example 13. Constructing a constraint graph from a 3-SAT-formula: φ = (A B C) (A D E) (D F G) }{{}}{{}}{{} v 1 v 2 1. Encode each clause as a vertex. v 3 2. Encode the satisfying assigments to a clause as the alphabet Σ. (T, T, T) (T, T, F) (T, F, T) (T, F, F) (F, T, T) (F, T, F) (F, F, T) Put a consistency constraint for every pair of clauses that a share a variable. 8
9 Theorem 14. Given a constraint graph G = (V, E), Σ, C with Σ 7, it is NP-hard to decide if SAT (G) = 1. Proof. Use the example from above to reduce to 3SAT. As 3SAT is NP-hard to decide, it is NP-hard to decide if SAT (G) = 1. To make it short, one can say a constraint graph is another data structure for boolean SAT-formulas. 4.2 Expander Graphs Simply speaking, an expander graph is a graph which expands quickly, which means that any subset of vertices is connected to many vertices of the complement set. This means expander graphs are graphs which are quite much random. To get these two sentences more formally, we will introduce a few definitions and a few examples. Definition 15. Let G = (V, E) be a d-regular graph. Let E(S, S) = (S S) E equal the number of edges from an non-empty subset S V to its complement. The edge expansion is defined as h(g) = min S, S V /2 E(S,S) S. To get an intuitive sight of the edge expansion, we should look at the following simple examples: Example 16. A disconnected graph has an expansion of 0. Proof. As the graph is disconnected, you can choose a connected component of the graph whose size is less than V /2 for S. Now E(S, S) = 0, and therefore h(g) = 0. 9
10 A random d-regular graph has an expansion of about d/2, independent of the number of vertices. Proof. Let S be a subset of at most n/2 vertices of a random d-regular graph. A typical of vertex in S is connected to d frac S n vertices in S. So E(S,S) S d frac S n. As S is minimal at about n/2, h(g) d/2. Lemma 17. There exist d 0 N and h 0 > 0, such that there is a polynomialtime constructible family {X n } n N of d 0 -regular graphs X n on n vertices with h(x n ) h 0. Instead of proving this lemma, we are instead presenting the following example for a family of expander graphs. Example 18. All graphs of size p (for all primes). Here V p = Z p and d = 3. Ever vertex is connected to its neighbors (x + 1,x 1) and its inverse (x 1 ). Working with expander graphs, one is especially interested in eigenvalues of the adjacency matrix of an expander G. These eigenvalues are also the Spectrum of the graph G. Some simple observations are that the largest eigenvalue of an expander is d or that isomorphic graphs have the same spectrum. Simply speaking, the largest eigenvalue tells you to how many vertices are connected to each other and the second largest eigenvalue tells you something about which vertices are connected to each other. This surely is very vague, but the next two definitions show what is meant exactly. Lemma 19. Let G be a d-regular graph, h(g) denotes the edge expansion of G and let λ(g) be the second largest eigenvalue of the adjacency matrix of G. Then λ(g) d h(g)2 d. This lemma shows the relation between the edge expansion h and the second largest eigenvalue. The following lemma shows how the second largest eigenvalue tells you something about the way vertices are connected. Lemma 20 (Expander Mixing Lemma). for all S, T V : E(S, T ) d S T λ S T If you look at the lemma, d S T n is exactly the number of edges you expect that connect S and T in a typical random graph. And E(S, T ) are the number of edges that connect S and T in the given graph. So the difference of those two terms stands for the difference of the given graph G to a typical random graph, this is obviously bound by λ. So, a small λ means a graph with allot of randomness. n 10
11 Theorem 21. Let G = (V, E) be a d-regular graph with a second largest eigenvalue λ. Let F E be a set of edges. The probability p that a random walk that starts at a random edge in F takes the i + 1st step in F as well, is bounded by F E + ( ) λ i. d Example 22 (Amplifying the success probability of random algorithms). L RP. A decides whether x L with m coin tosses and one-sided-error probability β. Simple way: Pr(A fails) β t and uses m t coin tosses. With random walk on expander graphs: Pr(A fails) (β + λ d )t and uses m + t log(d) coin tosses. 4.3 Probability Here we will shortly present a lemma from probability which will be used in the following presentation. Lemma 23. For any non-negative variable X, P r(x > 0) E2 (X) E(X 2 ). Proof. X is non-negative = E(X 2 ) = E(X 2 X > 0) P r(x > 0) and E(X) = E(X X > 0) P r(x > 0). = E2 (X) (E(X X>0) P r(x>0))2 E(X 2 ) = E(X 2 X>0) P r(x>0) P r(x > 0). because E(X 2 X > 0) E 2 (X X > 0). 4.4 Conclusion As this paper was written for JASS 06, I suggest to also read the paper [?] to get the complete picture of Probabilistically checkable proofs. References [Aro94] Sanjeev Arora. Probabilistic Checking of Proofs and Hardness of Approximation Problems. PhD thesis, CS Division, UC Berkeley, [Bel96] Mihir Bellare. Proof checking and approximation: Towards tight results. Sigact News, Vol.27, No. 1, [Din04] Irit Dinur. Advanced topics in complexity: Pcp theory, [Din05] Irit Dinur. The pcp theorem by gap amplification. Electronic Colloquium on Computational Complexity, [Has97] John Hastad. Some optimal inapproximability results. In the 28th Annual ACM Symposium on Theory of Computing, [HK97] Uri Zwick Howard Karloff. A 7/8-approximation algorithm for max 3sat?,
12 [NL03] Avi Wigderson Nati Linial. Expander graphs and their application, [Pap94] Christos Papadimitriou. Computational complexity. Addision-Wesley, [SH95] A. Steger S. Hougardy, H.J. Prömel. Probabilistically checkable proofs and their consequences for approximation algorithms. Trends in Discrete Mathematics, [Ves06] Bernhard Vesenmeyer. The pcp theorem by gap amplification, [Weg05] Ingo Wegener. Complexity theory : exploring the limits of efficient algorithms. Springer,
Notes for Lecture 2. Statement of the PCP Theorem and Constraint Satisfaction
U.C. Berkeley Handout N2 CS294: PCP and Hardness of Approximation January 23, 2006 Professor Luca Trevisan Scribe: Luca Trevisan Notes for Lecture 2 These notes are based on my survey paper [5]. L.T. Statement
More informationLecture 18: PCP Theorem and Hardness of Approximation I
Lecture 18: and Hardness of Approximation I Arijit Bishnu 26.04.2010 Outline 1 Introduction to Approximation Algorithm 2 Outline 1 Introduction to Approximation Algorithm 2 Approximation Algorithm Approximation
More informationPCPs and Inapproximability Gap-producing and Gap-Preserving Reductions. My T. Thai
PCPs and Inapproximability Gap-producing and Gap-Preserving Reductions My T. Thai 1 1 Hardness of Approximation Consider a maximization problem Π such as MAX-E3SAT. To show that it is NP-hard to approximation
More informationPCP Theorem and Hardness of Approximation
PCP Theorem and Hardness of Approximation An Introduction Lee Carraher and Ryan McGovern Department of Computer Science University of Cincinnati October 27, 2003 Introduction Assuming NP P, there are many
More informationEssential facts about NP-completeness:
CMPSCI611: NP Completeness Lecture 17 Essential facts about NP-completeness: Any NP-complete problem can be solved by a simple, but exponentially slow algorithm. We don t have polynomial-time solutions
More informationComplexity Classes IV
Complexity Classes IV NP Optimization Problems and Probabilistically Checkable Proofs Eric Rachlin 1 Decision vs. Optimization Most complexity classes are defined in terms of Yes/No questions. In the case
More informationLimits to Approximability: When Algorithms Won't Help You. Note: Contents of today s lecture won t be on the exam
Limits to Approximability: When Algorithms Won't Help You Note: Contents of today s lecture won t be on the exam Outline Limits to Approximability: basic results Detour: Provers, verifiers, and NP Graph
More informationCSC 5170: Theory of Computational Complexity Lecture 13 The Chinese University of Hong Kong 19 April 2010
CSC 5170: Theory of Computational Complexity Lecture 13 The Chinese University of Hong Kong 19 April 2010 Recall the definition of probabilistically checkable proofs (PCP) from last time. We say L has
More informationNon-Approximability Results (2 nd part) 1/19
Non-Approximability Results (2 nd part) 1/19 Summary - The PCP theorem - Application: Non-approximability of MAXIMUM 3-SAT 2/19 Non deterministic TM - A TM where it is possible to associate more than one
More informationLecture 16 November 6th, 2012 (Prasad Raghavendra)
6.841: Advanced Complexity Theory Fall 2012 Lecture 16 November 6th, 2012 (Prasad Raghavendra) Prof. Dana Moshkovitz Scribe: Geng Huang 1 Overview In this lecture, we will begin to talk about the PCP Theorem
More informationLecture 8 (Notes) 1. The book Computational Complexity: A Modern Approach by Sanjeev Arora and Boaz Barak;
Topics in Theoretical Computer Science April 18, 2016 Lecturer: Ola Svensson Lecture 8 (Notes) Scribes: Ola Svensson Disclaimer: These notes were written for the lecturer only and may contain inconsistent
More informationHardness of Approximation
CSE 594: Combinatorial and Graph Algorithms Lecturer: Hung Q. Ngo SUNY at Buffalo, Fall 2006 Last update: November 25, 2006 Hardness of Approximation 1 Overview To date, thousands of natural optimization
More information1 Probabilistically checkable proofs
CSC 5170: Theory of Computational Complexity Lecture 12 The Chinese University of Hong Kong 12 April 2010 Interactive proofs were introduced as a generalization of the classical notion of a proof in the
More information[PCP Theorem is] the most important result in complexity theory since Cook s Theorem. Ingo Wegener, 2005
PCP Theorem [PCP Theorem is] the most important result in complexity theory since Cook s Theorem. Ingo Wegener, 2005 Computational Complexity, by Fu Yuxi PCP Theorem 1 / 88 S. Arora, C. Lund, R. Motwani,
More informationU.C. Berkeley CS278: Computational Complexity Professor Luca Trevisan August 30, Notes for Lecture 1
U.C. Berkeley CS278: Computational Complexity Handout N1 Professor Luca Trevisan August 30, 2004 Notes for Lecture 1 This course assumes CS170, or equivalent, as a prerequisite. We will assume that the
More informationToday. Few Comments. PCP Theorem, Simple proof due to Irit Dinur [ECCC, TR05-046]! Based on some ideas promoted in [Dinur- Reingold 04].
Today Few Comments PCP Theorem, Simple proof due to Irit Dinur [ECCC, TR05-046]! Based on some ideas promoted in [Dinur- Reingold 04]. Remarkably simple novel proof. Leads to new quantitative results too!
More informationAdvanced Algorithms (XIII) Yijia Chen Fudan University
Advanced Algorithms (XIII) Yijia Chen Fudan University The PCP Theorem Theorem NP = PCP(log n, 1). Motivation Approximate solutions Definition (Approximation of MAX-3SAT) For every 3CNF formula ϕ, the
More informationKeywords. Approximation Complexity Theory: historical remarks. What s PCP?
Keywords The following terms should be well-known to you. 1. P, NP, NP-hard, NP-complete. 2. (polynomial-time) reduction. 3. Cook-Levin theorem. 4. NPO problems. Instances. Solutions. For a long time it
More informationCS154, Lecture 15: Cook-Levin Theorem SAT, 3SAT
CS154, Lecture 15: Cook-Levin Theorem SAT, 3SAT Definition: A language B is NP-complete if: 1. B NP 2. Every A in NP is poly-time reducible to B That is, A P B When this is true, we say B is NP-hard On
More informationTechnische Universität München Summer term 2010 Theoretische Informatik August 2, 2010 Dr. J. Kreiker / Dr. M. Luttenberger, J. Kretinsky SOLUTION
Technische Universität München Summer term 2010 Theoretische Informatik August 2, 2010 Dr. J. Kreiker / Dr. M. Luttenberger, J. Kretinsky SOLUTION Complexity Theory Final Exam Please note : If not stated
More informationCS Lecture 29 P, NP, and NP-Completeness. k ) for all k. Fall The class P. The class NP
CS 301 - Lecture 29 P, NP, and NP-Completeness Fall 2008 Review Languages and Grammars Alphabets, strings, languages Regular Languages Deterministic Finite and Nondeterministic Automata Equivalence of
More informationThe PCP Theorem by Gap Amplification
The PCP Theorem by Gap Amplification Irit Dinur February 13, 2007 Abstract The PCP theorem [3, 2] says that every language in NP has a witness format that can be checked probabilistically by reading only
More informationLecture 10: Hardness of approximating clique, FGLSS graph
CSE 533: The PCP Theorem and Hardness of Approximation (Autumn 2005) Lecture 10: Hardness of approximating clique, FGLSS graph Nov. 2, 2005 Lecturer: Venkat Guruswami and Ryan O Donnell Scribe: Ioannis
More informationThe Cook-Levin Theorem
An Exposition Sandip Sinha Anamay Chaturvedi Indian Institute of Science, Bangalore 14th November 14 Introduction Deciding a Language Let L {0, 1} be a language, and let M be a Turing machine. We say M
More informationDRAFT. PCP and Hardness of Approximation. Chapter 19
Chapter 19 PCP and Hardness of Approximation...most problem reductions do not create or preserve such gaps...to create such a gap in the generic reduction (cf. Cook)...also seems doubtful. The intuitive
More informationLecture 26. Daniel Apon
Lecture 26 Daniel Apon 1 From IPPSPACE to NPPCP(log, 1): NEXP has multi-prover interactive protocols If you ve read the notes on the history of the PCP theorem referenced in Lecture 19 [3], you will already
More informationProbabilistically Checkable Proofs and Approximating Solutions to Hard Problems
Probabilistically Checkable Proofs and Approximating Solutions to Hard Problems Kyle Littlefield June 8, 2005 Contents 1 Introduction 2 2 A Brief Introduction to Theory of Computation 3 2.1 Some Notation...........................
More informationApproximation Algorithms and Hardness of Approximation. IPM, Jan Mohammad R. Salavatipour Department of Computing Science University of Alberta
Approximation Algorithms and Hardness of Approximation IPM, Jan 2006 Mohammad R. Salavatipour Department of Computing Science University of Alberta 1 Introduction For NP-hard optimization problems, we
More informationComplexity Theory VU , SS The Polynomial Hierarchy. Reinhard Pichler
Complexity Theory Complexity Theory VU 181.142, SS 2018 6. The Polynomial Hierarchy Reinhard Pichler Institut für Informationssysteme Arbeitsbereich DBAI Technische Universität Wien 15 May, 2018 Reinhard
More informationOutline. Complexity Theory EXACT TSP. The Class DP. Definition. Problem EXACT TSP. Complexity of EXACT TSP. Proposition VU 181.
Complexity Theory Complexity Theory Outline Complexity Theory VU 181.142, SS 2018 6. The Polynomial Hierarchy Reinhard Pichler Institut für Informationssysteme Arbeitsbereich DBAI Technische Universität
More informationApproximation algorithm for Max Cut with unit weights
Definition Max Cut Definition: Given an undirected graph G=(V, E), find a partition of V into two subsets A, B so as to maximize the number of edges having one endpoint in A and the other in B. Definition:
More informationComplexity Classes V. More PCPs. Eric Rachlin
Complexity Classes V More PCPs Eric Rachlin 1 Recall from last time Nondeterminism is equivalent to having access to a certificate. If a valid certificate exists, the machine accepts. We see that problems
More information1 Agenda. 2 History. 3 Probabilistically Checkable Proofs (PCPs). Lecture Notes Definitions. PCPs. Approximation Algorithms.
CS 221: Computational Complexity Prof. Salil Vadhan Lecture Notes 20 April 12, 2010 Scribe: Jonathan Pines 1 Agenda. PCPs. Approximation Algorithms. PCPs = Inapproximability. 2 History. First, some history
More information-bit integers are all in ThC. Th The following problems are complete for PSPACE NPSPACE ATIME QSAT, GEOGRAPHY, SUCCINCT REACH.
CMPSCI 601: Recall From Last Time Lecture 26 Theorem: All CFL s are in sac. Facts: ITADD, MULT, ITMULT and DIVISION on -bit integers are all in ThC. Th The following problems are complete for PSPACE NPSPACE
More informationTopics in Complexity Theory
Topics in Complexity Theory Announcements Final exam this Friday from 12:15PM-3:15PM Please let us know immediately after lecture if you want to take the final at an alternate time and haven't yet told
More informationLecture 15: A Brief Look at PCP
IAS/PCMI Summer Session 2000 Clay Mathematics Undergraduate Program Basic Course on Computational Complexity Lecture 15: A Brief Look at PCP David Mix Barrington and Alexis Maciel August 4, 2000 1. Overview
More informationComplexity Theory. Jörg Kreiker. Summer term Chair for Theoretical Computer Science Prof. Esparza TU München
Complexity Theory Jörg Kreiker Chair for Theoretical Computer Science Prof. Esparza TU München Summer term 2010 Lecture 19 Hardness of Approximation 3 Recap Recap: optimization many decision problems we
More informationCS151 Complexity Theory. Lecture 15 May 22, 2017
CS151 Complexity Theory Lecture 15 New topic(s) Optimization problems, Approximation Algorithms, and Probabilistically Checkable Proofs Optimization Problems many hard problems (especially NP-hard) are
More informationGreat Theoretical Ideas in Computer Science
15-251 Great Theoretical Ideas in Computer Science Lecture 28: A Computational Lens on Proofs December 6th, 2016 Evolution of proof First there was GORM GORM = Good Old Regular Mathematics Pythagoras s
More informationNP and Computational Intractability
NP and Computational Intractability 1 Review Basic reduction strategies. Simple equivalence: INDEPENDENT-SET P VERTEX-COVER. Special case to general case: VERTEX-COVER P SET-COVER. Encoding with gadgets:
More informationIntro to Theory of Computation
Intro to Theory of Computation LECTURE 25 Last time Class NP Today Polynomial-time reductions Adam Smith; Sofya Raskhodnikova 4/18/2016 L25.1 The classes P and NP P is the class of languages decidable
More informationThe PCP Theorem by Gap Amplification
Electronic Colloquium on Computational Complexity, Revision 1 of Report No. 46 (2005) The PCP Theorem by Gap Amplification Irit Dinur September 25, 2005 Abstract We describe a new proof of the PCP theorem
More informationChapter 2. Reductions and NP. 2.1 Reductions Continued The Satisfiability Problem (SAT) SAT 3SAT. CS 573: Algorithms, Fall 2013 August 29, 2013
Chapter 2 Reductions and NP CS 573: Algorithms, Fall 2013 August 29, 2013 2.1 Reductions Continued 2.1.1 The Satisfiability Problem SAT 2.1.1.1 Propositional Formulas Definition 2.1.1. Consider a set of
More informationLecture 20: Course summary and open problems. 1 Tools, theorems and related subjects
CSE 533: The PCP Theorem and Hardness of Approximation (Autumn 2005) Lecture 20: Course summary and open problems Dec. 7, 2005 Lecturer: Ryan O Donnell Scribe: Ioannis Giotis Topics we discuss in this
More informationOn the Complexity of the Minimum Independent Set Partition Problem
On the Complexity of the Minimum Independent Set Partition Problem T-H. Hubert Chan 1, Charalampos Papamanthou 2, and Zhichao Zhao 1 1 Department of Computer Science the University of Hong Kong {hubert,zczhao}@cs.hku.hk
More informationThe PCP Theorem by Gap Amplification
Electronic Colloquium on Computational Complexity, Report No. 46 (2005) The PCP Theorem by Gap Amplification Irit Dinur April 16, 2005 Abstract Let C = {c 1,..., c n } be a set of constraints over a set
More informationIntroduction to Advanced Results
Introduction to Advanced Results Master Informatique Université Paris 5 René Descartes 2016 Master Info. Complexity Advanced Results 1/26 Outline Boolean Hierarchy Probabilistic Complexity Parameterized
More informationCS154, Lecture 13: P vs NP
CS154, Lecture 13: P vs NP The EXTENDED Church-Turing Thesis Everyone s Intuitive Notion of Efficient Algorithms Polynomial-Time Turing Machines More generally: TM can simulate every reasonable model of
More information1. Introduction Recap
1. Introduction Recap 1. Tractable and intractable problems polynomial-boundness: O(n k ) 2. NP-complete problems informal definition 3. Examples of P vs. NP difference may appear only slightly 4. Optimization
More informationLecture Hardness of Set Cover
PCPs and Inapproxiability CIS 6930 October 5, 2009 Lecture Hardness of Set Cover Lecturer: Dr. My T. Thai Scribe: Ying Xuan 1 Preliminaries 1.1 Two-Prover-One-Round Proof System A new PCP model 2P1R Think
More informationSAT, NP, NP-Completeness
CS 473: Algorithms, Spring 2018 SAT, NP, NP-Completeness Lecture 22 April 13, 2018 Most slides are courtesy Prof. Chekuri Ruta (UIUC) CS473 1 Spring 2018 1 / 57 Part I Reductions Continued Ruta (UIUC)
More informationA An Overview of Complexity Theory for the Algorithm Designer
A An Overview of Complexity Theory for the Algorithm Designer A.1 Certificates and the class NP A decision problem is one whose answer is either yes or no. Two examples are: SAT: Given a Boolean formula
More informationLecture 2 (Notes) 1. The book Computational Complexity: A Modern Approach by Sanjeev Arora and Boaz Barak;
Topics in Theoretical Computer Science February 29, 2016 Lecturer: Ola Svensson Lecture 2 (Notes) Scribes: Ola Svensson Disclaimer: These notes were written for the lecturer only and may contain inconsistent
More informationCMPT307: Complexity Classes: P and N P Week 13-1
CMPT307: Complexity Classes: P and N P Week 13-1 Xian Qiu Simon Fraser University xianq@sfu.ca Strings and Languages an alphabet Σ is a finite set of symbols {0, 1}, {T, F}, {a, b,..., z}, N a string x
More informationCSC 5170: Theory of Computational Complexity Lecture 9 The Chinese University of Hong Kong 15 March 2010
CSC 5170: Theory of Computational Complexity Lecture 9 The Chinese University of Hong Kong 15 March 2010 We now embark on a study of computational classes that are more general than NP. As these classes
More informationAlmost transparent short proofs for NP R
Brandenburgische Technische Universität, Cottbus, Germany From Dynamics to Complexity: A conference celebrating the work of Mike Shub Toronto, May 10, 2012 Supported by DFG under GZ:ME 1424/7-1 Outline
More informationTowards Optimal Lower Bounds For Clique and Chromatic Number
Towards Optimal Lower Bounds For Clique and Chromatic Number Lars Engebretsen 1, and Jonas Holmerin 2 1 MIT Laboratory for Computer Science 200 Technology Square, NE43-369 Cambridge, Massachusetts 02139-3594
More informationMore NP-Complete Problems
CS 473: Algorithms, Spring 2018 More NP-Complete Problems Lecture 23 April 17, 2018 Most slides are courtesy Prof. Chekuri Ruta (UIUC) CS473 1 Spring 2018 1 / 57 Recap NP: languages/problems that have
More informationReading group: proof of the PCP theorem
Reading group: proof of the PCP theorem 1 The PCP Theorem The usual formulation of the PCP theorem is equivalent to: Theorem 1. There exists a finite constraint language A such that the following problem
More informationDinur s Proof of the PCP Theorem
Dinur s Proof of the PCP Theorem Zampetakis Manolis School of Electrical and Computer Engineering National Technical University of Athens December 22, 2014 Table of contents 1 Introduction 2 PCP Theorem
More informationTheory of Computation CS3102 Spring 2014 A tale of computers, math, problem solving, life, love and tragic death
Theory of Computation CS3102 Spring 2014 A tale of computers, math, problem solving, life, love and tragic death Nathan Brunelle Department of Computer Science University of Virginia www.cs.virginia.edu/~njb2b/theory
More informationNP-Complete Problems. More reductions
NP-Complete Problems More reductions Definitions P: problems that can be solved in polynomial time (typically in n, size of input) on a deterministic Turing machine Any normal computer simulates a DTM
More informationLecture 6: Random Walks versus Independent Sampling
Spectral Graph Theory and Applications WS 011/01 Lecture 6: Random Walks versus Independent Sampling Lecturer: Thomas Sauerwald & He Sun For many problems it is necessary to draw samples from some distribution
More informationLecture 17 November 8, 2012
6.841: Advanced Complexity Theory Fall 2012 Prof. Dana Moshkovitz Lecture 17 November 8, 2012 Scribe: Mark Bun 1 Overview In the previous lecture, we saw an overview of probabilistically checkable proofs,
More information1.1 P, NP, and NP-complete
CSC5160: Combinatorial Optimization and Approximation Algorithms Topic: Introduction to NP-complete Problems Date: 11/01/2008 Lecturer: Lap Chi Lau Scribe: Jerry Jilin Le This lecture gives a general introduction
More informationOn the efficient approximability of constraint satisfaction problems
On the efficient approximability of constraint satisfaction problems July 13, 2007 My world Max-CSP Efficient computation. P Polynomial time BPP Probabilistic Polynomial time (still efficient) NP Non-deterministic
More information6.841/18.405J: Advanced Complexity Wednesday, April 2, Lecture Lecture 14
6.841/18.405J: Advanced Complexity Wednesday, April 2, 2003 Lecture Lecture 14 Instructor: Madhu Sudan In this lecture we cover IP = PSPACE Interactive proof for straightline programs. Straightline program
More informationApproximation Preserving Reductions
Approximation Preserving Reductions - Memo Summary - AP-reducibility - L-reduction technique - Complete problems - Examples: MAXIMUM CLIQUE, MAXIMUM INDEPENDENT SET, MAXIMUM 2-SAT, MAXIMUM NAE 3-SAT, MAXIMUM
More information1 Computational problems
80240233: Computational Complexity Lecture 1 ITCS, Tsinghua Univesity, Fall 2007 9 October 2007 Instructor: Andrej Bogdanov Notes by: Andrej Bogdanov The aim of computational complexity theory is to study
More informationPCP Soundness amplification Meeting 2 ITCS, Tsinghua Univesity, Spring April 2009
PCP Soundness amplification Meeting 2 ITCS, Tsinghua Univesity, Spring 2009 29 April 2009 Speaker: Andrej Bogdanov Notes by: Andrej Bogdanov Recall the definition of probabilistically checkable proofs
More informationComputational Complexity: A Modern Approach
1 Computational Complexity: A Modern Approach Draft of a book in preparation: Dated December 2004 Comments welcome! Sanjeev Arora Not to be reproduced or distributed without the author s permission I am
More informationNP-problems continued
NP-problems continued Page 1 Since SAT and INDEPENDENT SET can be reduced to each other we might think that there would be some similarities between the two problems. In fact, there is one such similarity.
More informationRestricted probabilistically checkable proofs. Jeffrey Finkelstein Computer Science Department, Boston University. May 5, 2015
Restricted probabilistically checkable proofs Jeffrey Finkelstein Computer Science Department, Boston University 1 Introduction May 5, 2015 In this work we provide some initial structural complexity results
More informationJASS 06 Report Summary. Circuit Complexity. Konstantin S. Ushakov. May 14, 2006
JASS 06 Report Summary Circuit Complexity Konstantin S. Ushakov May 14, 2006 Abstract Computer science deals with many computational models. In real life we have normal computers that are constructed using,
More informationLecture Notes Each circuit agrees with M on inputs of length equal to its index, i.e. n, x {0, 1} n, C n (x) = M(x).
CS 221: Computational Complexity Prof. Salil Vadhan Lecture Notes 4 February 3, 2010 Scribe: Jonathan Pines 1 Agenda P-/NP- Completeness NP-intermediate problems NP vs. co-np L, NL 2 Recap Last time, we
More informationAlgorithms Design & Analysis. Approximation Algorithm
Algorithms Design & Analysis Approximation Algorithm Recap External memory model Merge sort Distribution sort 2 Today s Topics Hard problem Approximation algorithms Metric traveling salesman problem A
More informationNP Complete Problems. COMP 215 Lecture 20
NP Complete Problems COMP 215 Lecture 20 Complexity Theory Complexity theory is a research area unto itself. The central project is classifying problems as either tractable or intractable. Tractable Worst
More informationRandomness and non-uniformity
Randomness and non-uniformity JASS 2006 Course 1: Proofs and Computers Felix Weninger TU München April 2006 Outline Randomized computation 1 Randomized computation 2 Computation with advice Non-uniform
More informationCS154, Lecture 18: 1
CS154, Lecture 18: 1 CS 154 Final Exam Wednesday December 12, 12:15-3:15 pm STLC 111 You re allowed one double-sided sheet of notes Exam is comprehensive (but will emphasize post-midterm topics) Look for
More informationIntroduction to Complexity Theory. Bernhard Häupler. May 2, 2006
Introduction to Complexity Theory Bernhard Häupler May 2, 2006 Abstract This paper is a short repetition of the basic topics in complexity theory. It is not intended to be a complete step by step introduction
More informationLecture 2. 1 More N P-Compete Languages. Notes on Complexity Theory: Fall 2005 Last updated: September, Jonathan Katz
Notes on Complexity Theory: Fall 2005 Last updated: September, 2005 Jonathan Katz Lecture 2 1 More N P-Compete Languages It will be nice to find more natural N P-complete languages. To that end, we ine
More informationThe Complexity of Optimization Problems
The Complexity of Optimization Problems Summary Lecture 1 - Complexity of algorithms and problems - Complexity classes: P and NP - Reducibility - Karp reducibility - Turing reducibility Uniform and logarithmic
More informationCS154, Lecture 13: P vs NP
CS154, Lecture 13: P vs NP The EXTENDED Church-Turing Thesis Everyone s Intuitive Notion of Efficient Algorithms Polynomial-Time Turing Machines More generally: TM can simulate every reasonable model of
More informationFormal definition of P
Since SAT and INDEPENDENT SET can be reduced to each other we might think that there would be some similarities between the two problems. In fact, there is one such similarity. In SAT we want to know if
More informationDesign and Analysis of Algorithms
Design and Analysis of Algorithms CSE 5311 Lecture 25 NP Completeness Junzhou Huang, Ph.D. Department of Computer Science and Engineering CSE5311 Design and Analysis of Algorithms 1 NP-Completeness Some
More information6.045: Automata, Computability, and Complexity (GITCS) Class 15 Nancy Lynch
6.045: Automata, Computability, and Complexity (GITCS) Class 15 Nancy Lynch Today: More Complexity Theory Polynomial-time reducibility, NP-completeness, and the Satisfiability (SAT) problem Topics: Introduction
More information2 P vs. NP and Diagonalization
2 P vs NP and Diagonalization CS 6810 Theory of Computing, Fall 2012 Instructor: David Steurer (sc2392) Date: 08/28/2012 In this lecture, we cover the following topics: 1 3SAT is NP hard; 2 Time hierarchies;
More informationCS278: Computational Complexity Spring Luca Trevisan
CS278: Computational Complexity Spring 2001 Luca Trevisan These are scribed notes from a graduate course on Computational Complexity offered at the University of California at Berkeley in the Spring of
More informationNP-problems continued
NP-problems continued Page 1 Since SAT and INDEPENDENT SET can be reduced to each other we might think that there would be some similarities between the two problems. In fact, there is one such similarity.
More informationINAPPROX APPROX PTAS. FPTAS Knapsack P
CMPSCI 61: Recall From Last Time Lecture 22 Clique TSP INAPPROX exists P approx alg for no ε < 1 VertexCover MAX SAT APPROX TSP some but not all ε< 1 PTAS all ε < 1 ETSP FPTAS Knapsack P poly in n, 1/ε
More informationComplexity Theory Final Exam
Technische Universität München Summer term 2010 Theoretische Informatik August 2, 2010 Dr. J. Kreiker / Dr. M. Luttenberger, J. Kretinsky Complexity Theory Final Exam Please note : If not stated otherwise,
More informationEasy Problems vs. Hard Problems. CSE 421 Introduction to Algorithms Winter Is P a good definition of efficient? The class P
Easy Problems vs. Hard Problems CSE 421 Introduction to Algorithms Winter 2000 NP-Completeness (Chapter 11) Easy - problems whose worst case running time is bounded by some polynomial in the size of the
More informationP is the class of problems for which there are algorithms that solve the problem in time O(n k ) for some constant k.
Complexity Theory Problems are divided into complexity classes. Informally: So far in this course, almost all algorithms had polynomial running time, i.e., on inputs of size n, worst-case running time
More informationSolution of Exercise Sheet 12
Complexity Theory (fall 2016) Solution of Exercise Sheet 12 Dominique Unruh Problem 1: Deterministic PCPs PCP stands for probabilistically checkable proofs. We will now show that the probabilistic part
More information2 Evidence that Graph Isomorphism is not NP-complete
Topics in Theoretical Computer Science April 11, 2016 Lecturer: Ola Svensson Lecture 7 (Notes) Scribes: Ola Svensson Disclaimer: These notes were written for the lecturer only and may contain inconsistent
More informationBCCS: Computational methods for complex systems Wednesday, 16 June Lecture 3
BCCS: Computational methods for complex systems Wednesday, 16 June 2010 Lecturer: Ashley Montanaro 1 Lecture 3 Hardness of approximation 1 Overview In my lectures so far, we ve mostly restricted ourselves
More informationNotes for Lecture Notes 2
Stanford University CS254: Computational Complexity Notes 2 Luca Trevisan January 11, 2012 Notes for Lecture Notes 2 In this lecture we define NP, we state the P versus NP problem, we prove that its formulation
More informationNP-Completeness I. Lecture Overview Introduction: Reduction and Expressiveness
Lecture 19 NP-Completeness I 19.1 Overview In the past few lectures we have looked at increasingly more expressive problems that we were able to solve using efficient algorithms. In this lecture we introduce
More informationPolynomial-time Reductions
Polynomial-time Reductions Disclaimer: Many denitions in these slides should be taken as the intuitive meaning, as the precise meaning of some of the terms are hard to pin down without introducing the
More information